Re: object algebra

"Neo" <neo55592_at_hotmail.com> wrote in message news:4b45d3ad.0402262127.953e296_at_posting.google.com...
> > > "Codd now regards nulls as an integral part of the relational model"
> >
> > So we have some disagreement in the relational community, at least with
> > Codd. However, Date is right, for the reasons in his book.
>
> Could you give Date's reason/conditions/logic, why he thinks NULLs are
> not integral part of RDM?

Read the chapter, as I'm too tired right now to give more than this summary (and messages pile up, so I doubt I'll get back to this one): he talks about missing information. He explains what it is, why it's difficult, and why nulls are a bad solution to the missing information problem: you quickly realize that null can mean many things, besides its implications for closure, and thus statements about nullable attributes quickly lose meaning as well. Special values, on the other hand, avoid the problem by being specific values VALID WITHIN A DOMAIN (aka type). How you define them is up to you (ALSO DEPENDS ON THE TYPE). Therefore (drumroll please) their meaning is clear. Received on Fri Feb 27 2004 - 22:11:10 CET